DE performance

Performance of test function

Searches for the matrix with the higher probability of having a CGCGCG primer (only one solution). Minimizes 1-p(CGCGCG)

test.perf <- read.csv('~/mnt/edann/hexamers/strand_specific/test_function.performance.csv', header = FALSE, col.names = c('it','score'))
head(test.perf)

Test of real function

Functions for reshaping and plotting are in DE_performance.r

max(long.mat$iter)

[1] "999"

l <- lapply(seq(1,3), function(i) plot.iteration(long.mat, i))
# png('~/AvOwork/output/DE_optimization/test_iter_matrix.png')
for(el in l){
  # png(paste0("~/AvOwork/output/DE_optimization/DE_iter-", el$data$iter[1], ".png"))
  plot(el)
  # dev.off()
}

MIN RHO OF DIFFERENT TESTS

With bigger population size I get better scores and if the algorithm gets stuck in a local minimum it happens after more iterations.

rho.files.10 <- list.files(opt.dir, pattern='_pop10_its300.DE.rho.txt', full.names = TRUE)
rho.files.20 <- list.files(opt.dir, pattern='_pop20_its300.DE.rho.txt', full.names = TRUE)
rho.files.30 <- list.files(opt.dir, pattern='_pop30_its300.DE.rho.txt', full.names = TRUE)
plot.optimization.score(rho.files.10) +
  ggtitle('Pop.size=10')

Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items

plot.optimization.score(rho.files.20) +
  ggtitle('Pop.size=20')

Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items

plot.optimization.score(rho.files.30) +
  ggtitle('Pop.size=30')

Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items
Read 300 items

Test on new optimization: maximizing density in region of interest

plot.optimization.score(score.ctcf.files) 

Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items

performance.mat <- read.csv(matrix.ctcf.files[4], row.names = 1)
long.mat <- reshape.prob.mat(performance.mat)

Using iter as id variables

head(performance.mat)

nrow(performance.mat)

[1] 100

lapply(matrix.ctcf.files, function(mat) 
  plot.iteration(reshape.prob.mat(read.csv(mat, row.names = 1)),99) +
    ggtitle(gsub(mat, pattern = '.+//', replacement = '')))

Using iter as id variables
Using iter as id variables
Using iter as id variables
Using iter as id variables
Using iter as id variables
Using iter as id variables
Using iter as id variables
Using iter as id variables
Using iter as id variables
Using iter as id variables

[[1]]

[[2]]

[[3]]

[[4]]

[[5]]

[[6]]

[[7]]

[[8]]

[[9]]

[[10]]

Test4 is so weird! Best score and ends with all Ts!!

performance.mat <- read.csv(matrix.ctcf.files[5], row.names = 1)
long.mat <- reshape.prob.mat(performance.mat)
lapply(c(1,10,20,50,70,99),function(it) plot.iteration(long.mat, it))

What does the profile full of Ts look like?

bestmat.bw <- import(bestmat.bw.file, format = 'BigWig')

Error in import(bestmat.bw.file, format = "BigWig") : 
  could not find function "import"

Average coverage on CTCF sites

plot.refpoint.profile.df(not.norm.bestmat.df)

[1]   0 300 600

plot.refpoint.profile.df(norm.bestmat.df)

[1]   0 300 600

Optimizing density of enriched kmers

Not total abundance

1. Popsize = 20

plot.optimization.score(score.ctcf.files) +
  theme(legend.title = element_blank())

Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items

2. Popsize = 30, 100 iterations

plot.optimization.score(score.ctcf.files) +
  theme(legend.title = element_blank())

Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items
Read 100 items

2. Popsize = 30, 100 iterations

plot.optimization.score(score.ctcf.files) +
  theme(legend.title = element_blank())

Read 200 items
Read 200 items
Read 200 items
Read 200 items
Read 200 items
Read 200 items
Read 200 items
Read 200 items
Read 200 items
Read 200 items

Load coverage profiles around CTCF sites

prof.files <- list.files(paste0(opt.dir,'predicted_cov_bestmat'), pattern = 'profile.txt', full.names = T)
profiles <- lapply(prof.files, load.profile)

Read 600 items
Read 600 items
Read 600 items
Read 600 items
Read 600 items
Read 600 items
Read 600 items

names(profiles) <- gsub(prof.files, pattern = '.+500reg|.bestMat.+profile.txt', replacement = '')
df <- make.df.of.profiles(profiles)
best.scores <- sapply(score.ctcf.files, function(x) rev.default(scan(x))[1])

Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items

names(best.scores) <- gsub(names(best.scores), pattern = '.+/|.DE.+.txt', replacement = '')
df.wScores <- df %>% mutate(score=best.scores[match(df$sample, names(best.scores))])
df.wScores %>% 
  filter(sample!='test1_ctcf_kmersFC_pop30_its100') %>% 
  filter(sample!="test10_ctcf_kmersFC_pop30_its100") %>% 
  plot.refpoint.profile.df(., color = 'score') +
  guides(color = guide_colorbar(barwidth = 20, barheight = 1))

[1]   0 300 600

Not normalized

profiles <- lapply(prof.files, load.profile, normalize=F)

Read 600 items
Read 600 items
Read 600 items
Read 600 items
Read 600 items
Read 600 items
Read 600 items

names(profiles) <- gsub(prof.files, pattern = '.+500reg|.bestMat.+profile.txt', replacement = '')
df <- make.df.of.profiles(profiles)
best.scores <- sapply(score.ctcf.files, function(x) rev.default(scan(x))[1])

Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items
Read 100 items
Read 100 items
Read 200 items

names(best.scores) <- gsub(names(best.scores), pattern = '.+/|.DE.+.txt', replacement = '')
df.wScores <- df %>% mutate(score=best.scores[match(df$sample, names(best.scores))])
df.wScores %>% 
  filter(sample!='test1_ctcf_kmersFC_pop30_its100') %>% 
  filter(sample!="test10_ctcf_kmersFC_pop30_its100") %>% 
  plot.refpoint.profile.df(., color = 'score') +
  guides(color = guide_colorbar(barwidth = 20, barheight = 1))

[1]   0 300 600

Look at the coverage profiles in a set of regions for the best matrixes

plot.expVSpred.coverage.track(l.best.bw[[2]])

The following non-numeric data column has been dropped: id

[[1]]
Genome axis 'Axis'

[[2]]
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr1
| positions: 4999
| samples:1
| strand: * 

regionsWctcf

 [1] "chr1 (1.82535e+08,1.8269e+08]"  "chr10 (8.91e+06,9.78e+06]"      "chr11 (7.0175e+07,7.032e+07]"   "chr13 (5.801e+07,5.816e+07]"   
 [5] "chr13 (9.9375e+07,9.9725e+07]"  "chr14 (2.1045e+07,2.111e+07]"   "chr14 (6.985e+07,6.989e+07]"    "chr15 (5.847e+07,5.878e+07]"   
 [9] "chr17 (2.6425e+07,2.6595e+07]"  "chr17 (2.776e+07,2.8335e+07]"   "chr17 (2.776e+07,2.8335e+07]"   "chr18 (6.0855e+07,6.1005e+07]" 
[13] "chr18 (6.4525e+07,6.461e+07]"   "chr18 (6.727e+07,6.7355e+07]"   "chr19 (3.2715e+07,3.305e+07]"   "chr19 (4.366e+07,4.4e+07]"     
[17] "chr2 (1.2636e+08,1.26775e+08]"  "chr2 (1.5342e+08,1.53505e+08]"  "chr4 (9.78e+06,9.88e+06]"       "chr6 (3.416e+07,3.426e+07]"    
[21] "chr6 (8.3135e+07,8.323e+07]"    "chr6 (1.48705e+08,1.49265e+08]" "chr7 (2.524e+07,2.552e+07]"     "chr7 (2.8335e+07,2.946e+07]"   
[25] "chr7 (2.8335e+07,2.946e+07]"    "chr7 (9.7675e+07,9.8015e+07]"   "chr7 (1.4244e+08,1.4268e+08]"   "chr8 (2.4525e+07,2.481e+07]"   
[29] "chr8 (5.366e+07,5.367e+07]"     "chr8 (8.4695e+07,8.47e+07]"     "chr8 (8.4695e+07,8.47e+07]"     "chr8 (1.0857e+08,1.08725e+08]" 
[33] "chr9 (1.06345e+08,1.06605e+08]" "chrX (1.53505e+08,1.5367e+08]" 

lapply(sample(regionsWctcf,10), function(x) plot.cov.wAnnotation(l.best.bw[x][[1]], ctcf))

The following non-numeric data column has been dropped: id

[[1]]
[[1]]$Axis
Genome axis 'Axis'

[[1]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr1
| positions: 5000
| samples:1
| strand: * 

[[1]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr1
| annotation features: 1

[[1]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[2]]
[[2]]$Axis
Genome axis 'Axis'

[[2]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr19
| positions: 4999
| samples:1
| strand: * 

[[2]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr19
| annotation features: 1

[[2]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[3]]
[[3]]$Axis
Genome axis 'Axis'

[[3]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr7
| positions: 4999
| samples:1
| strand: * 

[[3]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr7
| annotation features: 2

[[3]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[4]]
[[4]]$Axis
Genome axis 'Axis'

[[4]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr14
| positions: 4999
| samples:1
| strand: * 

[[4]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr14
| annotation features: 1

[[4]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[5]]
[[5]]$Axis
Genome axis 'Axis'

[[5]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr17
| positions: 4999
| samples:1
| strand: * 

[[5]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr17
| annotation features: 2

[[5]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[6]]
[[6]]$Axis
Genome axis 'Axis'

[[6]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr8
| positions: 4999
| samples:1
| strand: * 

[[6]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr8
| annotation features: 2

[[6]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[7]]
[[7]]$Axis
Genome axis 'Axis'

[[7]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr13
| positions: 4999
| samples:1
| strand: * 

[[7]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr13
| annotation features: 1

[[7]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[8]]
[[8]]$Axis
Genome axis 'Axis'

[[8]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr18
| positions: 4999
| samples:1
| strand: * 

[[8]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr18
| annotation features: 1

[[8]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[9]]
[[9]]$Axis
Genome axis 'Axis'

[[9]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr17
| positions: 4999
| samples:1
| strand: * 

[[9]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr17
| annotation features: 2

[[9]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[10]]
[[10]]$Axis
Genome axis 'Axis'

[[10]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr8
| positions: 4999
| samples:1
| strand: * 

[[10]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr8
| annotation features: 2

[[10]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 

Compare with even primer probability predicted coverage

plot.expVSpred.coverage.track(l.comon.bw[[2]])

The following non-numeric data column has been dropped: id

[[1]]
Genome axis 'Axis'

[[2]]
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr1
| positions: 4999
| samples:2
| strand: * 

lapply(sample(regionsWctcf,10), function(x) plot.cov.wAnnotation(l.comon.bw[x][[1]], ctcf))

The following non-numeric data column has been dropped: id

[[1]]
[[1]]$Axis
Genome axis 'Axis'

[[1]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr8
| positions: 4999
| samples:2
| strand: * 

[[1]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr8
| annotation features: 1

[[1]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[2]]
[[2]]$Axis
Genome axis 'Axis'

[[2]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr19
| positions: 4999
| samples:2
| strand: * 

[[2]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr19
| annotation features: 1

[[2]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[3]]
[[3]]$Axis
Genome axis 'Axis'

[[3]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr7
| positions: 4999
| samples:2
| strand: * 

[[3]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr7
| annotation features: 1

[[3]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[4]]
[[4]]$Axis
Genome axis 'Axis'

[[4]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr8
| positions: 4999
| samples:2
| strand: * 

[[4]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr8
| annotation features: 1

[[4]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[5]]
[[5]]$Axis
Genome axis 'Axis'

[[5]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr7
| positions: 4999
| samples:2
| strand: * 

[[5]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr7
| annotation features: 2

[[5]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[6]]
[[6]]$Axis
Genome axis 'Axis'

[[6]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr18
| positions: 4999
| samples:2
| strand: * 

[[6]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr18
| annotation features: 1

[[6]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[7]]
[[7]]$Axis
Genome axis 'Axis'

[[7]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr6
| positions: 4999
| samples:2
| strand: * 

[[7]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr6
| annotation features: 1

[[7]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[8]]
[[8]]$Axis
Genome axis 'Axis'

[[8]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr2
| positions: 5000
| samples:2
| strand: * 

[[8]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr2
| annotation features: 1

[[8]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[9]]
[[9]]$Axis
Genome axis 'Axis'

[[9]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr4
| positions: 4999
| samples:2
| strand: * 

[[9]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr4
| annotation features: 1

[[9]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 




[[10]]
[[10]]$Axis
Genome axis 'Axis'

[[10]]$`norm. coverage`
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr1
| positions: 5000
| samples:2
| strand: * 

[[10]]$CTCF
AnnotationTrack 'CTCF'
| genome: NA
| active chromosome: chr1
| annotation features: 1

[[10]]$titles
An object of class "ImageMap"
Slot "coords":
               x1        y1       x2        y2
Axis            6  37.24747 38.42335  77.20723
norm. coverage  6  77.20723 38.42335 278.26342
CTCF            6 278.26342 38.42335 318.47466

Slot "tags":
$title
            Axis   norm. coverage             CTCF 
          "Axis" "norm. coverage"           "CTCF" 

even.prof <- load.profile('~/mnt/edann/hexamers/strand_specific/artificial_coverage/highcov.random.42.artCov.CTCF.profile.txt')

Read 600 items

ctcf.prof <- load.profile('~/mnt/edann/hexamers/strand_specific/artificial_coverage/highcov.random.42test9_ctcf_kmersFC_pop30_its200.bestMat.coverage.artCov.CTCF.profile.txt')

Read 600 items

df <- make.df.of.profiles(list(
  CTCF=ctcf.prof,
  even=even.prof
))
plot.refpoint.profile.df(df)

[1]   0 300 600

Trying loading the whole matrix (halp)

tab <- read.table(gzfile('~/mnt/edann/hexamers/strand_specific/artificial_coverage/highcov.random.42test9_ctcf_kmersFC_pop30_its200.bestMat.coverage.artCov.CTCF.mat.gz'), skip = 1)
colnames(tab) <- c('chr', 'start', 'end', 'id', 'len', 'strand', seq(1,600))
tab.even <- read.table(gzfile('~/mnt/edann/hexamers/strand_specific/artificial_coverage/highcov.random.42.artCov.mat.gz'), skip = 1)
colnames(tab.even) <- c('chr', 'start', 'end', 'id', 'len', 'strand', seq(1,600))

tab %>% select(-c(chr,start,end,len, strand)) %>% 
  melt(id.vars=c('id'), variable.name='position', value.name='coverage') %>%
  group_by(id) %>% mutate(zscore=(coverage-mean(coverage, na.rm=T))/sd(coverage, na.rm=T)) %>%
  ungroup() %>% 
  group_by(position) %>% 
  summarise(avg=mean(zscore, na.rm=T)) %>%
  ggplot(., aes(as.numeric(position), avg)) +
    geom_line()

tab.even %>% select(-c(chr,start,end,len, strand)) %>% 
  melt(id.vars=c('id'), variable.name='position', value.name='coverage') %>%
  group_by(id) %>% mutate(zscore=(coverage-mean(coverage, na.rm=T))/sd(coverage, na.rm=T)) %>%
  ungroup() %>% 
  group_by(position) %>% 
  summarise(avg=mean(zscore, na.rm=T)) %>%
  ggplot(., aes(as.numeric(position), avg)) +
    geom_line()

EVEN COVERAGE

opt.dir <- '~/mnt/edann/hexamers/DEoptimization/even_cov/'
score.even.files.100 <- list.files(opt.dir, pattern = '.+even_pop30_its100.+rho.txt', full.names = TRUE)
score.even.files.200 <- list.files(opt.dir, pattern = '.+even_pop30_its200.+rho.txt', full.names = TRUE)
plot.optimization.score(score.even.files.100) +
  theme(legend.title = element_blank())
plot.optimization.score(score.even.files.200) +
  theme(legend.title = element_blank())

Variability in the initial population

original.pop.good <- read.csv(paste0(opt.dir, 'test5_even_pop30_its100.DE.originalPop.csv'), row.names = 1)
original.pop.bad <- read.csv(paste0(opt.dir, 'test1_even_pop30_its100.DE.originalPop.csv'), row.names = 1)

boxplot(as.vector(cor(t(original.pop.bad))), as.vector(cor(t(original.pop.good))), ylab='correlation between vectors', names=c('bad', 'good'))

Doesn’t seem to play a major role.

Visualizing in another way

originalpop.even.files <- list.files(opt.dir, pattern = '.+even_pop30_its200.+originalPop.csv', full.names = TRUE)
pops <- lapply(originalpop.even.files,function(x) read.csv(x, row.names = 1) %>% reshape.prob.mat %>% mutate(p=paste0(nuc,'.',pos), name=gsub(x, pattern = '.+/|.DE.+', replacement = '')) )
mean.pops <- do.call(rbind,pops)
mean.pops %>% group_by(name,nuc) %>% summarise(avg=mean(prob)) %>%
  ggplot(., aes(nuc, avg, group=name, color=name)) +
    # facet_grid(name~.) +
    geom_line() +
    # ylim(0,1) + 
    xlab('nucleotide')

This is interesting: less Cs in the ones that end up having a higher score

plot.expVSpred.coverage.track(l.comon.bw[[2]])

The following non-numeric data column has been dropped: id

[[1]]
Genome axis 'Axis'

[[2]]
DataTrack 'norm. coverage'
| genome: NA
| active chromosome: chr1
| positions: 4999
| samples:2
| strand: * 

lapply(sample(clean.common.bw,10), nice.plotTrack, labels=c('Even primer conc', 'Optimized primer conc'))

$`chr1 (5.7585e+07,5.7615e+07]`

$`chr10 (5.836e+07,5.837e+07]`

$`chr11 (2.6855e+07,2.686e+07]`

$`chr1 (9.763e+07,9.766e+07]`

$`chr10 (3.936e+07,3.938e+07]`

$`chr1 (5.5295e+07,5.5315e+07]`

$`chr1 (4.1015e+07,4.1045e+07]`

$`chr1 (1.82945e+08,1.82965e+08]`

$`chr10 (5.4115e+07,5.415e+07]`

$`chr10 (4.4135e+07,4.4185e+07]`

Looks good, more noisy than not optimized data.

Profile

ggsave('~/AvOwork/output/DE_optimization/even_optimization.pdf')

Saving 7 x 7 in image

Trying loading the whole matrix (halp)